4 ms·

> draft-schoen-intarea-unicast-lowest-address-00 Just as a heads up, zero in the last octet results in general reachability problems. We used x.y.z.0 as a vip

by dsp 5y ago

> draft-schoen-intarea-unicast-lowest-address-00 Just as a heads up, zero in the last octet results in general reachability problems. We used x.y.z.0 as a vip at Facebook^WMeta for a while, but learned the hard way that some people can't reach those ips. Some additional info from RIPE: https://labs.ripe.net/author/stephane_bortzmeyer/all-ip-addresses-are-equal-dot-zero-addresses-are-less-equal/ https://labs.ripe.net/author/stephane_bortzmeyer/all-ip-addr...

schoen 5y ago
Thanks! I've done my own testing more recently with the same RIPE Atlas infrastructure and the results were more encouraging. A good test for this is the AWS EC2 reachability test http://ec2-reachability.amazonaws.com/ http://ec2-reachability.amazonaws.com/ which puts many of the test hosts on .0 addresses. There was also a NANOG thread talking about this phenomenon. Decades-old standards do call for interoperability with a distant host whose address ends with a zero octet, so in theory there should be some interest toward cleaning this up on that basis alone. If you have some more data or examples or suggestions for where to find them, I'd be happy to hear about it.
- toast0 5y ago
  My anecdote is much older, but problems with reachability on .0 addresses was reported at Yahoo as well; I think .255s also had issues, so even in wider than /24 subnets, vips would not be assigned with last octet 0 or 255. From what I recall, it seemed to be CPE like modems and sometimes wireless routers that were blocking packets to (or sometimes from) the .0 and .255 addresses, usually under the guise of security; but it didn't seem to be ISP routers, or at least not often. RIPE Atlas can be useful for testing (my node just hit its 5th birthday!), but this might be something better tested with random clients. If you can convince someone running a speed test site to include a .0 test, that may be a way to get a little more coverage; or if you have budget, you might be able to make something work with web ads that beacons to a .0 and a .something else, and count the returns, etc. EDIT: I hadn't seen the EC2 reachability test; it's got some nice test addresses like 100.24.0.0 and 107.22.255.255, so that seems pretty comprehensive. It would be interesting if someone at AWS can share statistics on if there are significantly fewer images loaded from .0/.255 addresses than otherwise; and if they can narrow that down by AS or otherwise. Publicly would be great for my knowledge, but at least sharing with the ietf would be useful for decision making.
  - dsp 5y ago
    > it seemed to be CPE like modems and sometimes wireless routers that were blocking packets Agreed. I think D-Link devices were some of the most common we saw.

by dsp 5y ago

schoen 5y ago

Thanks! I've done my own testing more recently with the same RIPE Atlas infrastructure and the results were more encouraging. A good test for this is the AWS EC2 reachability test http://ec2-reachability.amazonaws.com/ http://ec2-reachability.amazonaws.com/ which puts many of the test hosts on .0 addresses. There was also a NANOG thread talking about this phenomenon. Decades-old standards do call for interoperability with a distant host whose address ends with a zero octet, so in theory there should be some interest toward cleaning this up on that basis alone. If you have some more data or examples or suggestions for where to find them, I'd be happy to hear about it.

toast0 5y ago
My anecdote is much older, but problems with reachability on .0 addresses was reported at Yahoo as well; I think .255s also had issues, so even in wider than /24 subnets, vips would not be assigned with last octet 0 or 255. From what I recall, it seemed to be CPE like modems and sometimes wireless routers that were blocking packets to (or sometimes from) the .0 and .255 addresses, usually under the guise of security; but it didn't seem to be ISP routers, or at least not often. RIPE Atlas can be useful for testing (my node just hit its 5th birthday!), but this might be something better tested with random clients. If you can convince someone running a speed test site to include a .0 test, that may be a way to get a little more coverage; or if you have budget, you might be able to make something work with web ads that beacons to a .0 and a .something else, and count the returns, etc. EDIT: I hadn't seen the EC2 reachability test; it's got some nice test addresses like 100.24.0.0 and 107.22.255.255, so that seems pretty comprehensive. It would be interesting if someone at AWS can share statistics on if there are significantly fewer images loaded from .0/.255 addresses than otherwise; and if they can narrow that down by AS or otherwise. Publicly would be great for my knowledge, but at least sharing with the ietf would be useful for decision making.
- dsp 5y ago
  > it seemed to be CPE like modems and sometimes wireless routers that were blocking packets Agreed. I think D-Link devices were some of the most common we saw.